CN108256429A - A kind of transmission tower object detection method using high spatial resolution single polarization SAR image - Google Patents
A kind of transmission tower object detection method using high spatial resolution single polarization SAR image Download PDFInfo
- Publication number
- CN108256429A CN108256429A CN201711376811.1A CN201711376811A CN108256429A CN 108256429 A CN108256429 A CN 108256429A CN 201711376811 A CN201711376811 A CN 201711376811A CN 108256429 A CN108256429 A CN 108256429A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- gradient
- transmission tower
- sar image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 49
- 230000005540 biological transmission Effects 0.000 title claims abstract description 24
- 230000010287 polarization Effects 0.000 title claims abstract description 6
- 238000000034 method Methods 0.000 claims abstract description 17
- 238000012937 correction Methods 0.000 claims abstract description 16
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 230000011218 segmentation Effects 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 3
- 230000001427 coherent effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007903 penetration ability Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
Description
方法领域method field
本发明涉及输电杆塔测技术领域，具体涉及一种利用高空间分辨率单极化SAR图像的输电杆塔目标检测方法。The invention relates to the technical field of transmission pole and tower measurement, in particular to a transmission pole and tower target detection method using high spatial resolution single-polarization SAR images.
背景方法background method
合成孔径雷达是一种主动式成像雷达，它能全天时地对目标进行观测，具有较高的分辨能力，遥感范围广且具有一定的穿透能力。随着遥感技术的不断发展，SAR技术被广泛地应用于军事和民用的多个领域。近年来，随着SAR图像数据的广泛获取并被普遍使用，SAR图像信息提取成为研究的热点。目标检测作为SAR图像信息提取中的关键技术环节也得到了蓬勃的发展。Synthetic aperture radar is a kind of active imaging radar, which can observe the target all the time, has high resolution ability, wide remote sensing range and certain penetration ability. With the continuous development of remote sensing technology, SAR technology is widely used in many fields of military and civilian use. In recent years, with the widespread acquisition and widespread use of SAR image data, SAR image information extraction has become a research hotspot. As a key technical link in SAR image information extraction, target detection has also been developed vigorously.
然而SAR图像由于受相干斑噪声的干扰，地物边界模糊，SAR图像的解译比较困难，如何抑制SAR图像中的相干斑噪声成为SAR图像目标检测技术的重要前提。面向对象的方法不仅能够有效抑制相干斑影响，还引入了更多可利用的特征，有助于理解图像所包含的地物及目标信息，增加目标检测的精度，同时降低检测结果中的虚警率。However, due to the interference of coherent speckle noise in SAR images, the boundaries of ground objects are blurred, and the interpretation of SAR images is difficult. How to suppress the coherent speckle noise in SAR images has become an important prerequisite for SAR image target detection technology. The object-oriented method can not only effectively suppress the influence of coherent speckle, but also introduce more available features, which is helpful to understand the ground objects and target information contained in the image, increase the accuracy of target detection, and reduce the false alarm in the detection results. Rate.
输电杆塔是电气设施中最重要的基础设施之一，其运行状态决定着整个电网的运行稳定和安全，对其进行目标检测具有重要意义。目前，在利用遥感影像进行输电杆塔的检测识别方面，国内外研究工作主要有利用多时相SAR图像变化检测的方法对电力塔进行形变监测，以及利用单时相高分辨率全极化SAR进行输电线路提取，对于实际应用范围更广的单一时相的单极化SAR数据，尚未提出一种利用高空间分辨率单极化SAR图像进行输电杆塔目标检测的方法。Transmission tower is one of the most important infrastructures in electrical facilities. Its operating status determines the stability and safety of the entire power grid, and its target detection is of great significance. At present, in the aspect of detection and identification of power transmission towers using remote sensing images, research work at home and abroad mainly includes the use of multi-temporal SAR image change detection methods to monitor the deformation of power towers, and the use of single-phase high-resolution full-polarization SAR for power transmission. For line extraction, for single-phase single-polarization SAR data with a wider range of practical applications, a method for detecting transmission tower targets using high-spatial-resolution single-polarization SAR images has not yet been proposed.
发明内容Contents of the invention
本发明的目的就是要针对现有技术的不足，提供一种利用高空间分辨率单极化SAR图像的输电杆塔目标检测方法，其通过计算和统计图像局部区域的梯度方向直方图来构成特征，增加了特征的维度，提高了目标检测的精度，进一步降低了目标检测中产生的虚警。The purpose of the present invention is to address the deficiencies of the prior art, to provide a transmission tower target detection method using a high spatial resolution single-polarization SAR image, which constitutes a feature by calculating and counting the gradient direction histogram of the local area of the image, The dimension of the feature is increased, the accuracy of target detection is improved, and the false alarm generated in target detection is further reduced.
为实现上述目的，本发明所涉及的一种利用高空间分辨率单极化SAR图像的输电杆塔目标检测方法，包括如下步骤：In order to achieve the above object, a method for detecting a power transmission tower target using high spatial resolution single-polarization SAR images involved in the present invention includes the following steps:
步骤1：根据原始的SAR图像，计算相对应的雷达功率图像；Step 1: Calculate the corresponding radar power image according to the original SAR image;
步骤2：根据SAR图像的功率图像进行多尺度分割，得到多尺度分割的结果；Step 2: Carry out multi-scale segmentation according to the power image of the SAR image, and obtain the result of multi-scale segmentation;
步骤3：结合多尺度分割的结果，使用面向对象的监督分类方法进行分类，得到初步检测结果；Step 3: Combining the results of multi-scale segmentation, use the object-oriented supervised classification method to classify, and obtain the preliminary detection results;
步骤4：根据SAR图像的功率图像，使用Gamma校正对输入图像进行标准化，得到标准化后的图像；Step 4: According to the power image of the SAR image, use Gamma correction to standardize the input image to obtain a standardized image;
步骤5：根据标准化后的图像，进行HOG特征计算；Step 5: Perform HOG feature calculation according to the standardized image;
步骤6：根据初步检测结果和图像的HOG特征，使用SVM分类器进行目标检测。Step 6: According to the preliminary detection results and the HOG features of the image, use the SVM classifier for target detection.
进一步地，所述步骤1中：Further, in the step 1:
第i个像元计算雷达功率图像的计算公式为：The formula for calculating the radar power image of the i-th pixel is:
式中：Realbi代表单极化SAR图像实部的像元值，Imagbi代表单极化SAR图像虚部的像元值，Pi代表第i个像元对应的功率图像像元值。In the formula: Real bi represents the pixel value of the real part of the single-polarization SAR image, Imag bi represents the pixel value of the imaginary part of the single-polarization SAR image, and P i represents the pixel value of the power image corresponding to the i-th pixel.
进一步地，所述步骤4中：Further, in the step 4:
所述Gamma校正的具体步骤为：The specific steps of the Gamma correction are:
步骤4.1：将功率图像进行归一化，将像素值转化为0和1之间的实数，得到归一化之后的图像；Step 4.1: Normalize the power image, convert the pixel value into a real number between 0 and 1, and obtain the normalized image;
步骤4.2：对归一化之后的图像进行预补偿，得到预补偿之后的图像；Step 4.2: precompensating the normalized image to obtain the precompensated image;
步骤4.3：对预补偿之后的图像进行反归一化，得到Gamma校正结果。Step 4.3: Denormalize the pre-compensated image to obtain a Gamma correction result.
作为优选项，所述步骤4.1中：As a preference, in the step 4.1:
所述第i个像元Pi归一化的计算公式为：The calculation formula of the normalization of the ith pixel Pi is:
式中：Pmax代表功率图像中像元的最大值，Pmin代表功率图像中像元的最小值，Ni代表归一化后i所对应的像元值。In the formula: P max represents the maximum value of the pixel in the power image, P min represents the minimum value of the pixel in the power image, and N i represents the pixel value corresponding to i after normalization.
作为优选项，所述步骤4.2中：As a preference, in the step 4.2:
所述对第i个像元预补偿的计算公式为：The calculation formula for the i-th pixel pre-compensation is:
式中，Ni为归一化之后的像元值，Gamma为预先确定的超参数，Yi为预补偿之后的输出值。In the formula, N i is the pixel value after normalization, Gamma is a predetermined hyperparameter, and Y i is the output value after pre-compensation.
作为优选项，所述步骤4.3中：As a preference, in the step 4.3:
所述第i个像元反归一化的计算公式为：The formula for calculating the denormalization of the i-th pixel is:
式中：Ymax代表预补偿图像中像元的最大值，Ymin代表预补偿图像中像元的最小值，P′i代表反归一化后i所对应的像元值。In the formula: Y max represents the maximum value of the pixel in the precompensated image, Y min represents the minimum value of the pixel in the precompensated image, and P′ i represents the pixel value corresponding to i after denormalization.
进一步地，所述步骤5中：Further, in the step 5:
所述计算HOG特征的具体步骤为：The specific steps for calculating the HOG feature are:
步骤5.1：计算Gamma校正后图像每个像素的梯度，包括大小和方向；Step 5.1: Calculate the gradient of each pixel of the image after Gamma correction, including size and direction;
步骤5.2：将图像划分成小Cell，统计每个Cell的梯度直方图；Step 5.2: Divide the image into small cells, and count the gradient histogram of each cell;
步骤5.3：将每几个Cell组成一个Block，并对Block的梯队直方图进行归一化，将所有Block的梯度直方图串联起来即得到图像的HOG特征。Step 5.3: Every few cells form a block, and the echelon histogram of the block is normalized, and the gradient histograms of all blocks are concatenated to obtain the HOG feature of the image.
作为优选项，所述步骤5.1中：As a preference, in the step 5.1:
所述计算Gamma校正后图像每个像素的梯度，具体的计算公式为：The gradient of each pixel of the image after the Gamma correction is calculated, and the specific calculation formula is:
对于图像中像素点(x，y)：For the pixel point (x, y) in the image:
Gx(x，y)＝H(x+1，y)-H(x-1，y)G x (x, y) = H (x+1, y) - H (x-1, y)
Gy(x，y)＝H(x，y+1)-H(x，y-1)G y (x,y)=H(x,y+1)-H(x,y-1)
式中:Gx(x，y)，Gy(x，y)，H(x，y)分别表示输入图像中像素点(x，y)处的水平方向梯度，垂直方向梯度和像素值。像素点(x，y)处的梯度幅值和梯度方向分别为：In the formula: G x (x, y), G y (x, y), H (x, y) respectively represent the horizontal direction gradient, vertical direction gradient and pixel value at the pixel point (x, y) in the input image. The gradient magnitude and gradient direction at the pixel point (x, y) are respectively:
作为优选项，所述步骤5.2中：As a preference, in the step 5.2:
将图像划分成小Cell，统计每个Cell的梯度直方图的具体步骤为：将图像按照一定的窗口大小划分为Cell，将每个Cell中的梯度幅值按照所在梯度方向区间进行加权投影，得到每个Cell的梯度直方图。The specific steps of dividing the image into small cells and counting the gradient histogram of each cell are as follows: divide the image into cells according to a certain window size, and perform weighted projection of the gradient amplitude in each cell according to the gradient direction interval, and obtain Gradient histogram of each Cell.
作为优选项，所述步骤5.3中：As a preference, in the step 5.3:
将每几个Cell组成一个Block，并对Block的梯队直方图进行归一化的具体步骤为：按照一定的步长，将几个Cell组成一个Block；对同一个Block内的所有梯度直方图进行归一化处理；完成所有Block的归一化，即得到整幅图像的梯度直方图。The specific steps to form a Block with several Cells and normalize the echelon histogram of the Block are: according to a certain step size, form several Cells into a Block; Normalization processing; complete the normalization of all Blocks, that is, obtain the gradient histogram of the entire image.
本发明的优点在于：其采用多尺度分割作为和面向对象监督分类的方法作为初步检测手段，既有效保留了地物目标的原始边界，又利用了单极化SAR图像中的统计信息，在抑制相干斑噪声的同时，提高了单极化SAR图像输电杆塔目标检测的准确性；使用HOG特征对初步检测结果进行再检测，通过计算和统计图像局部区域的梯度方向直方图来构成特征，增加了特征的维度，提高了目标检测的精度，进一步降低了目标检测中产生的虚警。The advantage of the present invention is that it adopts multi-scale segmentation and object-oriented supervised classification as the preliminary detection means, which not only effectively preserves the original boundaries of ground objects, but also utilizes the statistical information in single-polarization SAR images to suppress At the same time as coherent speckle noise, the accuracy of single-polarization SAR image transmission tower target detection is improved; HOG features are used to re-detect the preliminary detection results, and the features are formed by calculating and counting the gradient direction histogram of the local area of the image, which increases The dimensionality of features improves the accuracy of target detection and further reduces false alarms in target detection.
附图说明Description of drawings
图1为本发明的工作流程图；Fig. 1 is a work flow chart of the present invention;
图2为机载UAVSAR获得的L波段VV极化方式功率图；Figure 2 is the power diagram of the L-band VV polarization mode obtained by the airborne UAVSAR;
图3为多尺度分割得到的结果图；Figure 3 is the result map obtained by multi-scale segmentation;
图4为以分割结果和面向对象的监督分类得到的初步检测结果图；Figure 4 is a preliminary detection result diagram obtained by segmentation results and object-oriented supervised classification;
图5为机载L波段VV极化SAR图像的输电杆塔目标检测结果图。Figure 5 is a diagram of the target detection results of the transmission tower in the airborne L-band VV polarization SAR image.
具体实施方式Detailed ways
下面结合附图和具体实施例对本发明作进一步的详细描述：Below in conjunction with accompanying drawing and specific embodiment the present invention will be described in further detail:
如图1，一种利用高空间分辨率单极化SAR图像的输电杆塔目标检测方法，其包括如下步骤：As shown in Figure 1, a transmission tower target detection method using high spatial resolution single-polarization SAR images, which includes the following steps:
步骤1：根据原始的SAR图像，计算相对应的雷达功率图像：Step 1: According to the original SAR image, calculate the corresponding radar power image:
第i个像元计算雷达功率图像的计算公式为：The formula for calculating the radar power image of the i-th pixel is:
式中：Realbi代表单极化SAR图像实部的像元值，Imagbi代表单极化SAR图像虚部的像元值，Pi代表第i个像元对应的功率图像像元值。In the formula: Real bi represents the pixel value of the real part of the single-polarization SAR image, Imag bi represents the pixel value of the imaginary part of the single-polarization SAR image, and P i represents the pixel value of the power image corresponding to the i-th pixel.
步骤2：根据SAR图像的功率图像进行多尺度分割，得到多尺度分割的结果；Step 2: Carry out multi-scale segmentation according to the power image of the SAR image, and obtain the result of multi-scale segmentation;
步骤3：结合多尺度分割的结果，使用面向对象的监督分类方法进行分类，得到初步检测结果；Step 3: Combining the results of multi-scale segmentation, use the object-oriented supervised classification method to classify, and obtain the preliminary detection results;
步骤4：根据SAR图像的功率图像，使用Gamma校正对输入图像进行标准化，得到标准化后的图像：Step 4: According to the power image of the SAR image, use Gamma correction to standardize the input image to obtain the normalized image:
所述Gamma校正的具体步骤为：The specific steps of the Gamma correction are:
步骤4.1：将功率图像进行归一化，将像素值转化为0和1之间的实数，得到归一化之后的图像：Step 4.1: Normalize the power image, convert the pixel value into a real number between 0 and 1, and obtain the normalized image:
所述第i个像元Pi归一化的计算公式为：The calculation formula of the normalization of the ith pixel Pi is:
式中：Pmax代表功率图像中像元的最大值，Pmin代表功率图像中像元的最小值，Ni代表归一化后i所对应的像元值。In the formula: P max represents the maximum value of the pixel in the power image, P min represents the minimum value of the pixel in the power image, and N i represents the pixel value corresponding to i after normalization.
步骤4.2：对归一化之后的图像进行预补偿，得到预补偿之后的图像：Step 4.2: Pre-compensate the normalized image to obtain the pre-compensated image:
所述对第i个像元预补偿的计算公式为：The calculation formula for the i-th pixel pre-compensation is:
式中，Ni为归一化之后的像元值，Gamma为预先确定的超参数，Yi为预补偿之后的输出值。In the formula, N i is the pixel value after normalization, Gamma is a predetermined hyperparameter, and Y i is the output value after pre-compensation.
步骤4.3：对预补偿之后的图像进行反归一化，得到Gamma校正结果：Step 4.3: Denormalize the pre-compensated image to obtain the Gamma correction result:
所述第i个像元反归一化的计算公式为：The formula for calculating the denormalization of the i-th pixel is:
式中：Ymax代表预补偿图像中像元的最大值，Ymin代表预补偿图像中像元的最小值，P′i代表反归一化后i所对应的像元值。In the formula: Y max represents the maximum value of the pixel in the precompensated image, Y min represents the minimum value of the pixel in the precompensated image, and P′ i represents the pixel value corresponding to i after denormalization.
步骤5：根据标准化后的图像，进行HOG特征计算：Step 5: Perform HOG feature calculation based on the standardized image:
所述计算HOG特征的具体步骤为：The specific steps for calculating the HOG feature are:
步骤5.1：计算Gamma校正后图像每个像素的梯度，包括大小和方向：Step 5.1: Calculate the gradient of each pixel of the image after Gamma correction, including size and direction:
所述计算Gamma校正后图像每个像素的梯度，具体的计算公式为：The gradient of each pixel of the image after the Gamma correction is calculated, and the specific calculation formula is:
对于图像中像素点(x，y)：For the pixel point (x, y) in the image:
Gx(x，y)＝H(x+1，y)-H(x-1，y)G x (x, y) = H (x+1, y) - H (x-1, y)
Gy(x，y)＝H(x，y+1)-H(x，y-1)G y (x,y)=H(x,y+1)-H(x,y-1)
式中:Gx(x，y)，Gy(x，y)，H(x，y)分别表示输入图像中像素点(x，y)处的水平方向梯度，垂直方向梯度和像素值。像素点(x，y)处的梯度幅值和梯度方向分别为：In the formula: G x (x, y), G y (x, y), H (x, y) respectively represent the horizontal direction gradient, vertical direction gradient and pixel value at the pixel point (x, y) in the input image. The gradient magnitude and gradient direction at the pixel point (x, y) are respectively:
步骤5.2：将图像划分成小Cell，统计每个Cell的梯度直方图：Step 5.2: Divide the image into small cells, and count the gradient histogram of each cell:
将图像划分成小Cell，统计每个Cell的梯度直方图的具体步骤为：将图像按照一定的窗口大小划分为Cell，将每个Cell中的梯度幅值按照所在梯度方向区间进行加权投影，得到每个Cell的梯度直方图。The specific steps of dividing the image into small cells and counting the gradient histogram of each cell are as follows: divide the image into cells according to a certain window size, and perform weighted projection of the gradient amplitude in each cell according to the gradient direction interval, and obtain Gradient histogram of each Cell.
步骤5.3：将每几个Cell组成一个Block，并对Block的梯队直方图进行归一化，将所有Block的梯度直方图串联起来即得到图像的HOG特征：Step 5.3: Form every few cells into a block, and normalize the echelon histogram of the block, and concatenate the gradient histograms of all blocks to get the HOG feature of the image:
将每几个Cell组成一个Block，并对Block的梯队直方图进行归一化的具体步骤为：按照一定的步长，将几个Cell组成一个Block；对同一个Block内的所有梯度直方图进行归一化处理；完成所有Block的归一化，即得到整幅图像的梯度直方图。The specific steps to form a Block with several Cells and normalize the echelon histogram of the Block are: according to a certain step size, form several Cells into a Block; Normalization processing; complete the normalization of all Blocks, that is, obtain the gradient histogram of the entire image.
步骤6：根据初步检测结果和图像的HOG特征，使用SVM分类器进行目标检测。Step 6: According to the preliminary detection results and the HOG features of the image, use the SVM classifier for target detection.
本发明在实际使用时：When the present invention is actually used:
1.实例内容：1. Example content:
本发明实例实验的结果如图2～5所示。图2为机载UAVSAR获得的L波段VV极化方式功率图像，距离向和方位向分辨率均约为8m。图3为多尺度分割得到的结果；图4为以分割结果和面向对象的监督分类得到的初步检测结果；图5为本发明得到的检测结果。The results of the example experiments of the present invention are shown in Figures 2-5. Figure 2 is the power image of the L-band VV polarization mode obtained by the airborne UAVSAR, and the range and azimuth resolutions are both about 8m. Fig. 3 is the result obtained by multi-scale segmentation; Fig. 4 is the preliminary detection result obtained by segmentation result and object-oriented supervised classification; Fig. 5 is the detection result obtained by the present invention.
2.实验结果与分析：2. Experimental results and analysis:
从图2、图3、图4可以看出利用多尺度分割和面向对象的方法，大大的抑制了SAR图像中的相干斑噪声，更多的利用了SAR图像中的统计信息；而使用多级检测的手段，在提高检测率的同时，又进一步降低了检测结果中的虚警。It can be seen from Figure 2, Figure 3, and Figure 4 that the use of multi-scale segmentation and object-oriented methods can greatly suppress the coherent speckle noise in SAR images, and make more use of the statistical information in SAR images; while using multi-level The means of detection, while improving the detection rate, further reduces the false alarm in the detection results.
本发明采用多尺度分割的方法，较好的保留了目标的边界，利用面向对象的初步检测，成功抑制SAR图像中的相干斑噪声，降低了目标检测的难度，提取HOG特征后采用多级多特征的检测方式，在提高输电杆塔目标检测准确性的同时达到降低虚警的目的。The present invention adopts the method of multi-scale segmentation, which better preserves the boundary of the target, uses object-oriented preliminary detection, successfully suppresses the coherent speckle noise in the SAR image, and reduces the difficulty of target detection. The feature detection method can reduce false alarms while improving the detection accuracy of transmission tower targets.
最后，应当指出，以上实施例仅是本发明较有代表性的例子。显然，本发明不限于上述实施例，还可以有许多变形。凡依据本发明的方法实质对以上实施例所做的任何简单修改、等同变化及修饰，均应认为属于本发明的保护范围。Finally, it should be pointed out that the above embodiments are only representative examples of the present invention. Obviously, the present invention is not limited to the above-mentioned embodiments, and many variations are possible. Any simple modifications, equivalent changes and modifications made to the above embodiments according to the method of the present invention should be considered as belonging to the protection scope of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711376811.1A CN108256429A (en) | 2017-12-19 | 2017-12-19 | A kind of transmission tower object detection method using high spatial resolution single polarization SAR image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711376811.1A CN108256429A (en) | 2017-12-19 | 2017-12-19 | A kind of transmission tower object detection method using high spatial resolution single polarization SAR image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108256429A true CN108256429A (en) | 2018-07-06 |
Family
ID=62723667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711376811.1A Pending CN108256429A (en) | 2017-12-19 | 2017-12-19 | A kind of transmission tower object detection method using high spatial resolution single polarization SAR image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108256429A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109254287A (en) * | 2018-10-08 | 2019-01-22 | 国网经济技术研究院有限公司 | Method for detecting power transmission conductor based on synthetic aperture radar image |
| CN109359509A (en) * | 2018-08-27 | 2019-02-19 | 西安电子科技大学 | Polarimetric SAR ground object classification method based on OpenCL parallel |
| CN109376623A (en) * | 2018-10-08 | 2019-02-22 | 国网经济技术研究院有限公司 | Method for detecting transmission tower based on synthetic aperture radar image |
| CN110708511A (en) * | 2019-10-17 | 2020-01-17 | 山东浪潮人工智能研究院有限公司 | Monitoring video compression method based on image target detection |
| CN113313162A (en) * | 2021-05-25 | 2021-08-27 | 国网河南省电力公司电力科学研究院 | Method and system for detecting multi-scale feature fusion target |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104036530A (en) * | 2014-05-21 | 2014-09-10 | 中国科学院电子学研究所 | Processing method and device for SAR images |
| CN106407900A (en) * | 2016-08-31 | 2017-02-15 | 上海交通大学 | Multi-source aerial photo-based abnormal scene identification method |
| CN106651880A (en) * | 2016-12-27 | 2017-05-10 | 首都师范大学 | Method for detecting marine moving target of thermal infrared remote sensing image based on multi-feature fusion |
| CN106709463A (en) * | 2016-12-29 | 2017-05-24 | 武汉大学 | Hierarchical semantics-based SAR image classification method |
-
2017
- 2017-12-19 CN CN201711376811.1A patent/CN108256429A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104036530A (en) * | 2014-05-21 | 2014-09-10 | 中国科学院电子学研究所 | Processing method and device for SAR images |
| CN106407900A (en) * | 2016-08-31 | 2017-02-15 | 上海交通大学 | Multi-source aerial photo-based abnormal scene identification method |
| CN106651880A (en) * | 2016-12-27 | 2017-05-10 | 首都师范大学 | Method for detecting marine moving target of thermal infrared remote sensing image based on multi-feature fusion |
| CN106709463A (en) * | 2016-12-29 | 2017-05-24 | 武汉大学 | Hierarchical semantics-based SAR image classification method |
Non-Patent Citations (4)
| Title |
|---|
| 刘奕 等: "基于高分辨率ＳＡＲ影像的高压输电线路杆塔检测方法", 《电力科学与技术学报》 * |
| 蒋丹丹 等: "基于单极化TerraSAR_X影像提取建筑区研究", 《地理与地理信息科学》 * |
| 薛笑荣: "《SAR图像处理技术研究》", 31 August 2017, 科学技术文献出版社 * |
| 邓磊: "《SAR图像处理方法》", 28 February 2009, 测绘出版社 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109359509A (en) * | 2018-08-27 | 2019-02-19 | 西安电子科技大学 | Polarimetric SAR ground object classification method based on OpenCL parallel |
| CN109359509B (en) * | 2018-08-27 | 2021-09-28 | 西安电子科技大学 | OpenCL parallel-based polarized SAR terrain classification method |
| CN109254287A (en) * | 2018-10-08 | 2019-01-22 | 国网经济技术研究院有限公司 | Method for detecting power transmission conductor based on synthetic aperture radar image |
| CN109376623A (en) * | 2018-10-08 | 2019-02-22 | 国网经济技术研究院有限公司 | Method for detecting transmission tower based on synthetic aperture radar image |
| CN109376623B (en) * | 2018-10-08 | 2021-12-21 | 国网经济技术研究院有限公司 | Method for detecting transmission tower based on synthetic aperture radar image |
| CN110708511A (en) * | 2019-10-17 | 2020-01-17 | 山东浪潮人工智能研究院有限公司 | Monitoring video compression method based on image target detection |
| CN113313162A (en) * | 2021-05-25 | 2021-08-27 | 国网河南省电力公司电力科学研究院 | Method and system for detecting multi-scale feature fusion target |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108256429A (en) | A kind of transmission tower object detection method using high spatial resolution single polarization SAR image | |
| CN103400156B (en) | Based on the High Resolution SAR image Ship Detection of CFAR and rarefaction representation | |
| CN105023008B (en) | The pedestrian of view-based access control model conspicuousness and multiple features recognition methods again | |
| CN108268867B (en) | License plate positioning method and device | |
| CN109919002B (en) | Yellow stop line identification method and device, computer equipment and storage medium | |
| Ke et al. | SAR ship detection based on an improved faster R-CNN using deformable convolution | |
| CN106570863A (en) | Detection method and device for power transmission line | |
| US11854249B2 (en) | Character recognition method and terminal device | |
| CN111667655A (en) | Intrusion alarm device and method for high-speed railway safety zone based on infrared image | |
| CN108594226A (en) | A kind of mountain area SAR image power transmission tower frame detection method for taking landform into account | |
| CN105354824B (en) | DP-CFAR detection method based on extracted region | |
| CN104168462B (en) | Camera scene change detection method based on image angle point set feature | |
| CN105787470A (en) | Method for detecting power transmission line tower in image based on polymerization multichannel characteristic | |
| CN117333776A (en) | VOCs gas leakage detection method, device and storage medium | |
| CN104598922B (en) | Classification Method of Full Polarization SAR Image Based on Fuzzy C-Means | |
| CN105405138A (en) | Water surface target tracking method based on saliency detection | |
| CN109544608B (en) | Unmanned aerial vehicle image acquisition characteristic registration method | |
| CN104517124A (en) | SAR image change detection method based on SIFT feature point | |
| CN110136097A (en) | Insulator fault identification method and device based on characteristic pyramid | |
| Wu et al. | License plate location method based on texture and color | |
| Zeng et al. | Power transmission tower detection based on polar coordinate semivariogram in high-resolution SAR image | |
| CN115063381A (en) | Picture area change detection method and device, electronic equipment and storage medium | |
| CN102063726B (en) | Moving target classification method and system | |
| CN113298122B (en) | Target detection method and device and electronic equipment | |
| CN113250914A (en) | 2D displacement measurement method, system, equipment and storage medium for cabin tower top |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180706 |
|
| RJ01 | Rejection of invention patent application after publication |